Dietary habits across the primary-secondary school transition: A systematic review

Introduction: The transition to secondary school involves significant changes in children ’ s social and physical environment, which are often accompanied by changes in dietary habits. However, evidence around how dietary habits change during this life-stage transition is variable. Objective: This review aimed to identify, critically appraise, and summarise the wide-ranging evidence on changes in dietary habits across the primary – secondary school transition. Methods: Four electronic databases (PubMed, Embase, Web of Science, and Scopus) were searched for relevant studies published in English or Spanish from inception to February 2024. Longitudinal studies with a baseline quantitative assessment of dietary habits when children were in their last year or two of primary/elementary school with at least one follow-up assessment of dietary habits when children were in their first or second year of secondary/middle/junior-high school were eligible for inclusion. Results: Seven studies (10 independent samples) were eligible for inclusion, and fourteen different dietary habits outcomes were reported across the seven studies. The frequency of consuming breakfast, fruit, vegetables, and milk decreased across the primary-secondary school transition. Inconsistent results were found regarding changes in the frequency of consumption of sugar-sweetened beverages, fruit juice, snacks, and fast-foods. Many dietary habits were only evaluated in single studies. Conclusions: There is moderate evidence for a decrease in the frequency of consumption of breakfast, fruit, vegetables, and milk in children across the transition from primary to secondary school. This highlights this timeframe as a valuable period for intervention to promote the continuation of positive dietary habits. More high-quality longitudinal studies using homogeneous methodologies to evaluate changes in dietary habits across the school transition are needed to fully understand the implications of this transition on children ’ s diet and health behaviours.


Introduction
Unhealthy dietary patterns and behaviours during childhood and early adolescence can contribute to the development of obesity and high adiposity (Dalwood et al., 2020;de Menezes et al., 2023), as well as other adverse health consequences such as lower bone mineral density (de Menezes et al., 2023), higher incidence of dental caries (Mahboubi et al., 2021), higher blood pressure, higher risk for metabolic syndrome (Dalwood et al., 2020), poorer mental health and lower quality of life (Dalwood et al., 2020;Lane et al., 2022;Wu et al., 2019).Childhood and adolescence are key periods for the acquisition of healthy dietary habits, as unfavourable dietary habits during this period have been previously associated with poorer diet quality in adulthood (Hu et al., 2016;Lipsky et al., 2017).National and international dietary guidelines suggest that dietary patterns adequate for optimal health should be characterised by a high consumption of fruits and vegetables (at least 5 portions per day) and starchy foods (especially higher fibre or wholegrain varieties), alongside limited intake of foods that are high in fat, salt, and sugar (National Health Service, 2022;Sociedad Española de Nutrición Comunitaria, 2018;Harvard T.H. Chan School of Public Health, 2011).However, unfavourable dietary patterns are highly prevalent in young people.In the UK for example, only 11% of adolescents between 11 and 18 years meet the recommendations of 5 servings of fruits and vegetables per day, and a high percentage consume foods high in sugars, fats, and salt (Public Health England, 2021).Furthermore, 45% of UK adolescents are habitual consumers of sugar-sweetened beverages (SSB), 48% are habitual consumers of cakes and pastries, and 60% are habitual consumers of crisps and savoury snacks (Public Health England, 2021).
To inform the development of interventions capable of promoting favourable dietary habits and demoting unfavourable dietary habits in young people, it is necessary to understand the factors that influence specific dietary habits.Environmental influences on young people's dietary habits are critical (Gubbels, 2020).The food environments that children operate in are overwhelmed with access and opportunities for children to choose and consume cheap, energy-dense, nutrient poor foods over nutrient dense foods (Gebremariam et al., 2015;Neufeld et al., 2022).Children spend half of their waking hours within the school environment (Driessen et al., 2014) and schools are increasingly recognised as a significant influence on young people's dietary habits (Gubbels, 2020).The dietary habits of primary/middle school aged children are heavily influenced by the behaviours and decision making of their parents and carers (Gubbels, 2020), whereas transitioning to secondary school means significant changes in the social and physical environment that are accompanied by increased autonomy around decisions for purchasing foods, which are often influenced by peers, increased access and opportunities for foods, higher exposure to social media and advertisements for unhealthy foods (Haerens et al., 2010;Neufeld et al., 2022;Story et al., 2009).The transition from childhood to adolescence is therefore often characterised by changes in dietary habits, such as a decrease in fruit and vegetables consumption, increases in sugar sweetened beverages and energy-dense snacks (Gebremariam et al., 2015), which makes it a key time for public health intervention to improve child/adolescent health.
Systematic reviews have started to examine both the tracking (stability) and changes in individual energy balanced-related health behaviours across the primary-secondary school transition.Reviews have shown that sedentary behaviours increase (Chong et al., 2020;Pearson et al., 2017) and physical activity decreases (Chong et al., 2020) across this transition, whereas evidence around dietary habits seem more variable (Emke et al., 2023).Therefore, the aim of this review was to identify, critically appraise and summarise the wide-ranging, recent evidence on changes in dietary habits across the primarysecondary school transition in an attempt to provide updated consensus on the existing evidence in this area.

Methods
This systematic review was presented in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) (Liberati, 2009), and the review protocol was registered with the International Prospective Register for Systematic Reviews (PROSPERO: registration number CRD42023425234).

Search strategy
Searches of electronic databases (PubMed, Embase, Web of Science, and Scopus) were conducted between May 2023 and January 2024.The search strategy was developed around groups of keywords: dietary habits, transitions, study type, and population.Supplementary searches were conducted that included manual searches of personal files, and screening reference lists of articles.The search syntax was first developed for PubMed and then adapted to the database-specific search requirements.Search strategy and syntax developed for PubMed are provided in Supplementary File 1.No date limitations were applied to the searches.

Inclusion and exclusion criteria
For inclusion, studies were required to (1) use a longitudinal observational design with at least two points of assessment, (2) measured children's dietary habits during the last year or two of primary/ elementary school (e.g., aged 9-11 years), and measured dietary habits of these children again in the first year or two of secondary/middle/ junior-high school (e.g., 11-13 years); (3) include a measure and a quantitative outcome of dietary habit at both (or all) points of assessment; (4) be published in a peer-reviewed journal in the English or Spanish language up to and including January 2024.

Identification of relevant studies
Potentially relevant studies were identified by (1) screening the titles; (2) screening the abstracts; (3) if abstracts were not available or did not provide sufficient data, the entire article was retrieved and screened.Full texts of the potentially relevant studies were screened by two independent reviewers (APS, NP) to determine whether they met the inclusion criteria, and any discrepancies were resolved through discussion with a third reviewer (EH).

Data extraction
Data were extracted on structured forms developed for this review.Extracted data included author, date of publication and country, participant characteristics (sample age at baseline and follow-up, gender, and sample size), length of follow-up, dietary behaviours assessed, method of dietary behaviour measurement, type of analysis and subgroup analyses.This information is summarised in Table 1.

Study quality assessment
Included studies were assessed for methodological quality using the National Institute of Health (NIH) quality assessment tool for Observational Cohort and Cross-Sectional Studies (National Institutes of Health, 2018).The tool includes fourteen quality items, six of them were not applicable for the aims of this systematic review and were therefore removed.Eight quality items were included related to the following dimensions: clearly stated research question, clearly specified study population, representative sample and justified sample size, non-biased recruitment of subjects, valid and reliable assessment tool, adequate follow-up rate (see Supplementary File 2) for complete description of items used).
Each quality item was scored as 'yes', 'no', 'cannot be determined', 'not applicable' or 'not reported'.Two researchers (APS, FB) independently assessed each study against the 8 study quality items and any discrepancies were resolved through discussion.When consensus could not be reached, a third reviewer was consulted (NP).The positive scores ('yes') were summed and converted to a percentage, which indicated the overall quality of the article.Based on these percentages, a study was considered 'high quality' when scoring 80-100%, 'moderate quality' when scoring 40-79% and 'poor quality' when scoring ≤39%.

Data synthesis
Data on individual dietary habit stability and/or change over the primary-secondary school transition were extracted from independent samples (e.g.boys and girls) within the included articles and summarised in Tables 2 and 3 respectively.In public health and epidemiological literature, tracking is used to describe the (relative) stability of a certain characteristic over time (Twisk, 2002).Tracking coefficients (r) were extracted from included articles and were classified as small (0.10-0.29), moderate (0.30-0.49) or large (≥0.5) according to strength of association cut-off points described by Cohen (Cohen, 2013).

Study characteristics
After removing duplicates, the literature search yielded 8277 potentially relevant articles (see Fig. 1), of which eight articles were eligible for inclusion in this review (see Table 1).These eight articles included seven studies, as two articles used data from the same study (Dowda et al., 2017;Taverno Ross et al., 2020), and ten independent samples, as three studies considered boys and girls separately (Chen et al., 2005;Cullen & Zakeri, 2004;Dowda et al., 2017).Studies were conducted in the USA (n = 4) (Cullen & Zakeri, 2004;Dowda et al., 2017;Lytle et al., 2000;Oza-Frank et al., 2012;Taverno Ross et al., 2020), Japan (n = 1) (Chen et al., 2005), Australia (n = 1) (Marks et al., 2015), and Belgium (n = 1) (Vereecken et al., 2010).Studies were published between 2000 and 2020.All studies reported dietary behaviours for boys and girls combined, and three studies (37.5%) (Chen et al., 2005;Cullen & Zakeri, 2004;Dowda et al., 2017;Taverno Ross et al., 2020) also reported dietary behaviours for boys and girls separately.In         all studies, the first point of assessment was conducted in the last two years of primary/elementary school (4th -6th grades, depending on the country), when participants were aged from 9 to 12 years.Follow-up periods ranged from 5 months to 3 years, and one study included a second follow-up during the first two years of secondary school (Dowda et al., 2017).All studies used self-reported questionnaires (completed by children) to assess dietary behaviours.Fourteen different dietary behavioural outcomes were reported across the seven included studies.
All studies reported on changes in dietary behaviours, and two reported on tracking (stability) over the time (Chen et al., 2005;Taverno Ross et al., 2020;Vereecken et al., 2010).All seven studies were rated as moderate quality according to the National Institute of Health (NIH) quality assessment tool for Observational Cohort and Cross-Sectional Studies (National Institutes of Health, 2018) (Table 1).

Stability of dietary behaviours across the primary-secondary school transition
The findings from the two studies that reported on stability of dietary behaviours across the primary-secondary school transition are presented in Table 2.Both studies assessed stability over three years (Chen et al., 2005;Vereecken et al., 2010).Stability of breakfast eating was small (Boys: r = 0.23; Girls: r = 0.26) (Chen et al., 2005), while stability of fruit and vegetable consumption, and excess score (calculated by combining SSB, sweets and crisps consumption, with a higher score indicating greater consumption of these foods) was moderate (r = 0.38 and r = 0.43, respectively) (Vereecken et al., 2010).

Changes in dietary behaviours across the primary-secondary school transition
Table 3 summarises the findings related to changes in dietary behaviours across the transition from elementary/primary school to secondary/middle school.The overall trend for change in dietary behaviours was a decrease in the consumption of fruit, vegetables, and milk, as well as a decrease in breakfast frequency.Findings regarding changes in intake of SSB, energy-dense snacks and fast-food were inconsistent.
Changes in breakfast frequency were assessed in two studies (Chen et al., 2005;Lytle et al., 2000), where the proportion of children eating breakfast decreased by 2% and 9% over the primary-secondary school transition, respectively.
Five studies found a decrease in fruit consumption (daily or during school/outside school hours) (Cullen & Zakeri, 2004;Lytle et al., 2000;Marks et al., 2015;Oza-Frank et al., 2012;Taverno Ross et al., 2020), and four studies found a decrease in vegetable consumption (daily or during school/non-school hours) across the school transition (Cullen & Zakeri, 2004;Lytle et al., 2000;Oza-Frank et al., 2012;Taverno Ross et al., 2020).One study found no change in daily vegetable consumption over the time, but a decrease in vegetable consumption during school authors; -= significant decreases in dietary behaviour; -= decreases in dietary behaviour but non-significant or statistical significance data has not been provided by authors.hours (Marks et al., 2015).One study found an increase in the consumption of specific types of vegetables (potatoes and green salads) (Cullen & Zakeri, 2004;Oza-Frank et al., 2012).Two studies (Marks et al., 2015;Vereecken et al., 2010) found a decrease in consumption of fruit and vegetables combined (Marks et al., 2015;Vereecken et al., 2010).
Two studies reported a decrease in energy-dense snack consumption (i.e., sweet snacks, salty snacks, non-core foods) over the transition (Lytle et al., 2000;Marks et al., 2015), and one study reported an increase in consumption of French fries (Cullen & Zakeri, 2004).One study found a decrease in fast-food consumption across the school transition (Oza-Frank et al., 2012), and one study found an increase of the percentage of adolescents who ate fast food, but a decrease in the frequency of eating fast food across the school transition (Lytle et al., 2000).
Other domains of dietary behaviours were assessed in single studies.Decreases were found in diet quality (Dowda et al., 2017;Taverno Ross et al., 2020), eating lunch (Lytle et al., 2000), excess score (combining on SSB, sweets and crisps consumption) (Vereecken et al., 2010), and protein consumption (Taverno Ross et al., 2020).Increases were found in the frequency of eating dinner (Lytle et al., 2000) and in wholegrains score (based on the consumed quantity) (Taverno Ross et al., 2020).There were no significant changes over the time for consumption of dairy (Taverno Ross et al., 2020) or the healthiness of children's food choices from their school canteen (Marks et al., 2015).

Differences by sex or ethnicity in dietary habits change
Two studies reported data on changes in dietary behaviours over the time by sex or ethnicity, finding similar tendencies in changes in dietary behaviours over the across transition among sexes and different ethnic groups (Cullen & Zakeri, 2004;Dowda et al., 2017).Additionally, one study found a positive association between being a girl and change in FV score (β = 0.312; SE = 0.073) (Vereecken et al., 2010).

Differences by country in dietary habits change
When considering the domains for which results vary between studies, the study conducted in Australia (Marks et al., 2015) found a decrease over the time in SSB and energy-dense snacks consumption, whereas the studies conducted in the USA showed varying results regarding SSB, with two of them (Cullen & Zakeri, 2004;Lytle et al., 2000) finding an increase in SSB consumption over the time, while one of them found a decrease (Oza-Frank et al., 2012).Studies from USA also found inconsistencies in energy-dense snack consumption, with one reporting an increase in consumption (Cullen & Zakeri, 2004), while other one found a decrease in energy-dense snack consumption across the primary-secondary school transition (Oza-Frank et al., 2012).

Discussion
This systematic review provides the most current and comprehensive synthesis of evidence of change in dietary behaviours across the transition from primary to secondary school.Across seven studies (eight papers), and 10 independent samples, findings of changes in dietary behaviours were mixed, due in part to high heterogeneity of the measurement and outcome of dietary behaviours reported.Moreover, few studies considered changes in dietary behaviours according to important sociodemographic characteristics such as sex or ethnicity, which limits our ability to make conclusions on whether tracking or change in diet across the school transition is more or less apparent in certain subgroups.
Notable findings include evidence of reduced fruit, vegetable, milk, and breakfast intake over the primary-secondary school transition period.Reductions in fruit and vegetable intake are concerning as these are foods that are already poorly consumed in children (Boushey et al., 2020).Previous longitudinal studies have also found a decrease in the frequency of consumption of fruits and vegetables over time, albeit from age 11-13 years and not focusing on the school transition (Gebremariam et al., 2013;van der Sluis et al., 2010).Also, fewer secondary schools comply with the school food standards compared to primary schools, with evidence that unhealthier food is more commonly available throughout the school day, and unhealthy snacks are often used as treats/rewards and for fundraising events (McIntyre et al., 2022).This fact, combined with the low availability of fruits and vegetables present in many schools (Murray et al., 2015) and increases in individuals' autonomy around food choices occurring alongside the school transition (Neufeld et al., 2022), may contribute to the observed reduction in fruit and vegetable consumption at the transition to secondary school.Previous studies suggest a trend of declining dietary behaviours (such as a decrease in fruit and vegetable consumption) from early to late adolescence (Doggui et al., 2021).In this sense, it is important to highlight that in the present results, the only study showing no change in daily vegetable consumption had the lowest follow-up period (5-8 months) and showed a decrease in consumption during school hours (Marks et al., 2015).Such evidence underscores the importance of further work to facilitate increases in fruit and vegetables intake in children during the beginning of their transition to secondary schools, as well as the need to improve the access to fruits and vegetables in secondary school environments to ensure tracking of nutritious behaviours.This review also found convincing evidence that the frequency of eating breakfast decreases across the primary-school transition.It is possible that the transition to secondary school, and the changes in schedules and distance to school for example, may contribute to chaotic and rushed mornings where breakfast is easily skipped (Harvey-Golding et al., 2015).Furthermore, as children become adolescents their diurnal sleep patterns change which could result in them not feeling hungry first thing in the morning (Colrain & Baker, 2011).In addition, increases in autonomy, the increased influence of peers and media associated with starting at secondary school, as well as changes to the food accessible and available to purchase and consume, likely all contribute to individuals making unhealthier dietary choices, such as skipping breakfast (Driessen et al., 2014;Neufeld et al., 2022).Given the importance of breakfast for cognitive performance and academic achievement during childhood and adolescence (e.g., Lundqvist et al., 2019), further research should aim to understand the key determinants of changes in breakfast consumption as children transition to adolescence, as well as the focus on changing school environments, to inform the development of public health interventions to promote sustained healthy breakfast habits.There is also the need for greater understanding of the breakfast foods being consumed as young people transition from primary to secondary school due to most of the studies (including those in this review) simply evaluating whether breakfast is consumed or not, rather than the type, quality and quantity of foods consumed (Adolphus et al., 2016).
Milk consumption was also shown to reduce over the school transition, which could be related to the reduction in breakfast consumption, as milk is commonly consumed during breakfast (Rampersaud et al., 2005), or due to milk not commonly being offered to children in secondary schools.Evidence regarding SSB intake was variable.The two studies that reported an increase in the consumption of SSB over the time were carried out in the USA and these results are similar to those from another longitudinal study in the USA including 780 children who increased their SSB from 11 to 13 years of age (Wiecha et al., 2006).
Á. Peral-Suárez et al.Given the links between SSB intake with obesity, cardiometabolic ill-health, dental issues, and some cancers (Malik & Hu, 2022;Valenzuela et al., 2021), these findings underscore the need for work to continue to focus on reducing children's intake of SSBs and suggest that the school transition period could be a valuable time for intervention to support adolescents' healthful choices.
For non-core foods (energy dense snacks and fast-food) again the results varied without remarkable differences between countries.There was some evidence showing a decrease in consumption of non-core foods over the school transition, while some studies showed an increase in consumption.The inconsistent results found between studies, despite in all cases there being a change in the school and thus food environment associated with going to secondary school, could be explained by findings from other studies, suggesting that adolescent attitudes, peer modeling (i.e.doing what friends do), and intentions were more strongly associated with the consumption of SSB and energy dense snacks than the physical school environment (van der Horst et al., 2008).Evidence for the changes in many dietary behaviours was captured in single studies which prevents conclusions from being made, but it is worth noting that we found evidence of a decrease in overall diet quality, a change in the type of foods consumed, and changes in the consumption of meals across the school transition.
Alongside the findings of the current review, it is important to consider the broader context of changes in children's eating behaviours and diet quality that occur during the transition from childhood to adolescence (Mikkilä et al., 2005).There is evidence that the quality of dietary intake decreases from childhood to adolescence (da Costa et al., 2024), and that children who are low consumers of foods such as fresh fruit, high fibre breakfast cereals, high fibre bread, potatoes (boiled) and legumes at age 7 years are likely to remain low consumers at age 13 years (Ambrosini et al., 2014), indicating stability of less healthy dietary behaviours.While the results of the current review are in line with these broader developmental observations, the context of changing school environment appears to be an important influence on changes in the consumption of certain foods.A study by Marks et al., 2015 sought to examine whether moving from primary/middle school to a completely new school environment for secondary/high school had a greater or lesser impact on eating behaviours than remaining within the same school environment.Students who changed school environments for secondary education reported less healthy dietary behaviours such as higher purchasing of snack foods, a greater decrease in fruit and vegetables consumption during school hours, and a smaller decrease in SSB consumption (Marks et al., 2015).Such findings suggest that changing school environment plays a role in the change of certain dietary behaviours, with changes to the food environment, the desire to fit in with new peers, and more independence around eating and food choices likely combining to drive changes in eating behaviours as children transition to secondary schools (Ashton, 2008;Brown et al., 2017).Further research is required to disentangle the impact of changing school environments on specific eating behaviours from the broader developmental changes that we see in eating behaviours.More research comparing the eating behaviours of children who change school environments with the eating behaviours of those who remain within the same school environment is needed.Furthermore, a better understanding of the determinants of the changes in eating behaviours and patterns over the school transition is essential to underpin strategies to prevent a decline in healthier food consumption and to help school environments become more conducive to healthier food choices in both primary and secondary schools.
The stability of the dietary behaviours examined within this review (breakfast, fruit and vegetable consumption, and SSB, sweets and crisps together (i.e. the excess score)) was moderate (ranging from 0.38 to 0.43) across durations of follow-up (2-3 years), with the exception of breakfast which showed a small stability, indicating that the breakfast eating behaviour changes from baseline to follow-up.However, future studies assessing the stability of behaviours across the primary-secondary school transition in different subgroups of children are essential to identifying potential differences in trajectories according to these subgroups, and to developing effective interventions in both primary and secondary school environments, with the aim of promoting the maintenance of adequate consumption of healthy foods within the context of a varied and balanced diet.
Strengths of this review include the comprehensive search strategy within four electronic databases without publication date restrictions, and the double screening by two independent researchers of titles/abstracts and quality assessment resulting in the elimination of bias and errors in the methodology.A further strength is this review's focus on the time periods closely surrounding the school transition such that it focused on articles reporting data from the last two years of primary school as baseline data and then on the first two years of secondary school at follow-up.The inclusion of such studies provides a more accurate reflection of any changes in behaviour around this important transition period.
However, there are limitations, some of which are due to the literature itself.The studies included in the present review were of moderate quality with no high-quality studies identified.Future work using standardised measures would facilitate more conclusive synthesis and comparison of studies examining dietary changes during the school transition.Furthermore, due to the inclusion criteria, it is possible that studies which did not clearly state that measurements were taken from children attending either of the last two years of a primary school and from the same children when they were attending the first two years of a secondary school were missed or excluded because we focused our evidence synthesis on explicit school transitions implying a change in the school environment.Due to the heterogeneity in methodology and reporting of outcomes, conducting a meta-analysis was not feasible or appropriate.The studies adopted a diverse range of measures for assessing dietary behaviours, and data were all reported by children, both of which makes drawing reliable conclusions more difficult.Also, only three of the included studies specified the type of school (public or private) where the data collection was carried out, and these three were conducted in public schools.Therefore, differences in the stability or changes in dietary habits between those attending public or private schools could not be identified from this review.Due to the low number of studies that met the inclusion criteria, most continents are represented within just one study which means that differences between different geographical areas may be poorly represented.

Conclusions
This review makes a novel contribution to the literature by focusing on changes in a wide range of dietary behaviours with a specific focus on the last two years of primary school and first two years of secondary school.The current systematic review found moderate evidence for a decrease in breakfast, fruit, vegetable, and milk consumption of children/adolescents across the transition from primary to secondary school which highlights this timeframe as a valuable period for intervention to promote ongoing positive dietary behaviours.There was inconclusive evidence for other dietary behaviours due to inconsistent methodologies, results just from single studies, and a lack of high-quality studies.More longitudinal studies using homogeneous methodologies to evaluate changes in dietary behaviours across the school transition are needed, specifically studies from different countries, high-quality studies which control for potential sources of bias, and studies where analyses are conducted to highlight differences according to socioeconomic groups.

75%a
Included data form more than one cohort, only data from cohort 1 met the inclusion criteria, and was used in this review; ES = elementary school, MS = middle school, HS = high school, PS = primary school, SS = secondary school; BG = boys and girls assessed together, B/G = boys and girls assessed separately, G = Girls, B=Boys, OWB = overweight boys, OWG = overweight girls; F = frequency, HF = high fat, SSB = sugar-sweetened beverages, HEI = Healthy Eating Index, FV = fruit and vegetables, C = consumption, FFQ = food frequency questionnaire, SES = socio-economic status.

Table 1
Study characteristics.

Table 2
Dietary behaviour tracking coefficients, by length of follow-up and behaviours assessed.

Table 3
Change in dietary behaviour across the primary-secondary school transition.